The START Trial: On the Shoulders of SMART

1. The START Trial: On the Shoulders of SMART 5 years after SMART INSIGHT Satellite Session WAC, Washington DC, July 2012

2. Rationale for START • Considerable variation in time from infection to progression to CD4 count < 350 cells/µL • Lack of data from randomized controlled trials - and conflicting findings from observational studies - to assess benefit:risk ratio for ART in early HIV infection • ART reduces infectiousness and public health would argue for early ART • Pathogenesis studies suggest potential benefit of suppression of viral replication • ART associated with potential side effects, risk of resistance and challenge of long term adherence

3. Hypothesis: ART reduces risk of non-AIDS events by dampening inflammation and coagulation No ART: risk* inflammation coagulation activation HIV ART: risk* *: magnitude of absolute risk or depends on demographics, lifestyle, co-morbidities, host genetic

4. The case why START may show net harm from early ART • Incidence of morbidity and mortality in early HIV without ART low • In particular among younger persons • Overrepresented among those with early HIV • ART may adversely affect • Kidney function • Bone formation • Liver function • Cardiovasvular disease • Neurocognitive function • Pulmonary function • (Cancer ?) • If scenario turns out to correct (will know in next 3 yrs) – major implications for persons already started early ART

5. First, do no harm • Primum non nocere • The doctor should not prescribe medications unless s/he knows that the treatment is unlikely to be harmful Doctor oath, year 1200

6. Areas of consensus and of controversy • Consensus • Continued substantial transmission remains Achilles heel in overcoming the HIV pandemic • Initiation of ART makes persons less infectious • ART is of net healthbenefit to HIV+ persons with HIV-related symptoms or with CD4<350 cells/µL • Controversy • Is ART of net healthbenefit to the asymptomatic HIV+ person if started at a CD4 count>350 cells/µL (i.e. early ART) ?

7. Recent guidelines favouring use of early ART • US DHHS guidelines, March 2012 • Argue that early ART is of benefit to the HIV+ person = all should betreated • Quality of evidence: < 350 (AI), 350-500 (AII), > 500 (BIII) • WHO Couples HIV testing and counselling, April 2012 • Recognises uncertainty in benefit:risk ratio from early use of ART for the HIV+ person • Recommend that ART is offered to prevent transmission in heterosexual couples as long as • the HIV+ person is told that there is no evidence of benefit – and there maybe harm - to his/her health • [assume also: the person makes the decision autonomously and without pressure from his/her partner]    

8. RCTs to inform ”earlier start” of ART • Focus interpretation ondeferralstrategy • Median CD4 count and outliers (=thoseallowed to progress to evenlower CD4 counts) • 3 completedRCT’s – • Haiti trial (NEJM 2010): 280 versus 166 • SMART subgroup (JID 2008): 477 versus 236 • HPTN 052 (NEJM 2011 + IAS 2012): 442 versus 225 • In START • - Current : 644 vs deferred ? median CD4 in earlyvsdeferredgroup:

9. GRADE criteria for assessing evidence:used widely and adopted by WHO • ”Observational studies […] only yield low quality evidence; in unusual circumstances […] classified as moderate/high. E.g. if [..] magnitude of treatment effect provides extremely large and consistent estimates” • Whether to start ART in early HIV infection • 3 observational studies • When to start collaboration (Lancet 2009), NA ACCORD (NEJM 2009), Causalcollaboration (Ann Intern Med, 2011) • Results not large nor consistent (RH close to 1 in two and in one > 2) • GRADE classification: C (A is best) While SMART wasconducted: SeveralsmallersizeRCT’s and observational studies suggestedthatinterruption of ART wassafe ! (http://www.gradeworkinggroup.org/FAQ/index.htm)

10. Survival after ART initiated at different CD4 count levels between200-500 cells/µL: ”causal” modelling Proportion surviving Absolute risk differences: Death: -0.02% AIDS/death: 2.1% The HIV-CAUSAL Collaboration, Ann Intern Med 2011

11. START Design HIV-infected adults, ART-naive with CD4+ cell counts > 500 cells/mm3 Early ART Group Immediately initiate ART N=2,000 Deferred ART Group Defer ART until CD4+ <350 cells/mm3 or symptoms develop N=2,000 Primaryendpoint: Serious AIDS & seriousnon-AIDSdisease (375) Current Status: 2700randomised; randomisation ends <2012/ study <2015.

12. START Design HIV-infected adults, ART-naive with CD4+ cell counts > 500 cells/mm3 235 sites in 31 countries and involving 1000+ experienced HIV clinicians 100-140 persons randomised / month This reflects clinical equipoise in clinical research community globally Early ART Group Immediately initiate ART N=2,000 Deferred ART Group Defer ART until CD4+ <350 cells/mm3 or symptoms develop N=2,000 Primaryendpoint: Serious AIDS & seriousnon-AIDSdisease (375) Current Status: 2700randomised; randomisation ends <2012/ study <2015.

13. Keycurrentstatistics of START • N=2693 – enrollment complete in 1Q2013 • Demographics • Age=35 (38% above 40 yrs), 19% females • 56% white, 22% black • 65% MSM, 31% heterosexual – 1 year since HIV diagnosis • Clinical status • 5% HCV and 2% HBV co-infected • CD4 count = 644 (IQR:580-748); HIV-RNA = 14,110 • 11% TDR (of 699 tested sofare) • 37% smokers, 20% hypertension • Median follow-up: 13 months (38% > 18); 3.1% LTFU • 8 substudies (organ dysfunction (pulmonary, neurology, CVD, bone) + host gene & informed consent)

14. Importance of defining balance of risk: benefit to individuals versus prevention in resourceconstrained regions • No observational data on risk:benefit of early ART (to date, all published to date from high resource countries) • Limited data from HPTN 052 • Deferral strategy lead to initiation of ART @ CD4 count considerably below current WHO recommendations (350) • composite clinical endpoint, • benefit restricted to extrapulmonary TB, • no benefit in pulmonary TB (surprising !) • no benefit on survival

15. Thank you